Decouple the replacement zone scan from PERFORMANCE_MODE

[liamiak]

PERFORMANCE_MODE currently gates two unrelated behaviours:

site	what it does	documented?
Spell.java:101	skips the LKI re-control when activator != controller	yes — this is what the settings text describes and warns about
ReplacementHandler.java:110	the Tap/Untap zone skip from #11160	no — not mentioned in the UI at all

The settings description reads:

Disables additional static abilities checks to speed up the game engine. (Warning: breaks some 'as if had flash' scenarios when casting cards owned by opponents).

That describes only the first. And it's accurate — skipping the re-control means checkActivatorRestrictions evaluates "You" against c.getController(), which is still the opponent, so a MayPlay permission granted to you doesn't match. That's the reported Petty Larceny behaviour (its exiled cards become unplayable, lands included).

So the second behaviour — a large, rules-derived win — is hidden behind a checkbox users are correctly told will break their cards. Not everyone enables it, which means #11160 is inert for most users and never runs in CI.

Measurements

3 headless 4-player games, "Big 240531" mirror, normalised per call (sims aren't deterministic, so game length varies between runs):

	before	after
Tap scan	2788.6 µs	194.4 µs
ProduceMana scan	2749.2 µs	178.7 µs
RE-list rebuilds per scan	1681	34

For comparison, on the same workload the Spell.canPlay branch fires on 25.8% of calls at 161.8 µs each — about 3.8s, against ~140s for the scan restriction. The safe half is roughly 97% of the benefit.

What this changes

The zone restriction becomes unconditional and derives its zones from the loaded card pool rather than a hardcoded pair. Per CR 113.6 these replacements only function from the battlefield (or the command zone, for Effects); a card that needs one active elsewhere declares ActiveZones$, and that widens the scan for that event.

PERFORMANCE_MODE keeps its current default and now governs only the Spell.canPlay branch — exactly what its description says.

It also closes a gap

The current skip returns before replacementEffect.zonesCheck(cardZone), so it ignores ActiveZones$ entirely. A custom card that declared ActiveZones$ Graveyard on a Tap replacement would still be skipped — the flag's stated purpose ("in case a custom card wants one active from elsewhere") isn't actually served today; it's all-or-nothing. Verified by adding such a card and watching the computed set:

stock pool:  Tap=[Battlefield, Command]
+ probe:     Tap=[Graveyard, Battlefield, Command]     (Untap/ProduceMana unchanged)

Notes

• Includes unit tests for the zone derivation (defaults, declaration widening, event isolation, non-hot events unrestricted).
• The traversal itself is unchanged; this only avoids the per-card getReplacementEffects() rebuild, which measured as ~78% of scan cost. Restricting the traversal is a possible follow-up.
• Supersedes AI: include ProduceMana in the performance-mode replacement zone skip #11327, which extends the hardcoded list to ProduceMana. If you'd rather take the small version first, that one stands alone and I'll rebase this on top.
• Happy to instead keep the restriction behind a new, separately-named flag if you'd prefer an escape hatch for undeclared custom cards — I went with unconditional because the declaration mechanism already exists and makes the flag redundant, but it's your call.

AI credit: Claude (opus 4.8) is doing a the work here.

Edit/note from non-AI: I'm mainly pushing back on things that don't quite make sense to me (this one took a bit of settling to get to something that seemed an actual fix instead of symptom remediation (which could rhyme with bloat).

[tool4ever]

interesting idea, but at the same time pretty over-engineered
especially if it's only for something easily hardcoded
the right approach is just getting rid of the overhead instead of trying to detect it: #11058

[liamiak]

On #11058 being the right approach here — one profiling note, in case it's useful for sequencing. I profiled one cantHappenCheck(Tap) on a 2-player board of ~560 cards — so 560 iterations of the visitor loop — and split where the ~465µs goes:

• the traversal itself (forEachCardInGame walking the zones, empty body): ~4.5µs, ~1%
• c.getReplacementEffects() rebuilding each card's effect list inside the loop: ~370µs, ~79%

forEachCardInGame is already allocation-free by construction, and #11058 as drawn streams that traversal — so it targets the 1%. The 79% is the per-card getReplacementEffects() rebuild, which a stream iterates the same way; flatMap over the zones still calls it once per card. So streaming this entry point shouldn't move this path much on its own.

The thing that would actually retire the skip — for every event, not just these three — is making that per-card rebuild cheap: either caching the RE list per card the way keywords already are (getCachedKeywords/updateKeywordsCache), or a per-event replacement index like activeTriggers. Both are bigger than this and I don't think they're on a branch yet; the LKI/preList path in getReplacementList makes a naive per-card cache tricky (it computes effects against a hypothetical battlefield state, not the live card). So I don't think this contradicts the #11058 direction — just that this particular scan's cost is somewhere the traversal rework won't reach, that I can see.

[tool4ever]

Ok, thanks for checking
that means #11323 might not be needed and I need to take a look at #11314 too